Tootfinder

Day 8 (a bit late): Timnit Gebru
Academic authors are authors too, and there are a bunch of people I deeply respect both in my fields and adjacent.
Gebru is someone I have huge respect for because she stood up for her (mild, completely reasonable) principles to the point of losing her job on Google's AI ethics team (since disbanded entirely), and then went ahead and founded an independent research institute to continue doing AI ethics research.
Why was she fired? Because she insisted on publishing her "Stochastic Parrots" paper after it passed Google internal review only to have extra nonstandard scrutiny applied at the last minute. Why did Google want to suppress her paper (which included an academic co-author)? Because it expressed valid criticisms of the large language models fad, and Google was planning to make money off that fad. Personally, I don't think I'd hire an "AI ethics" team only to then try to suppress their publications, and Google seems to now agree, having scrapped the team (during the initial furor, Timnit's boss also effectively quit to support her).
That "Stochastic Parrots" paper? Indeed, it predicts the core underlying problems with large language models that lead to so many of their user-side harms today. You can read it here: #20AuthorsNoMen

Few-Shot Neuro-Symbolic Imitation Learning for Long-Horizon Planning and Acting
Pierrick Lorang, Hong Lu, Johannes Huemer, Patrik Zips, Matthias Scheutz
https://arxiv.org/abs/2508.21501

Few-Shot Neuro-Symbolic Imitation Learning for Long-Horizon Planning and Acting
Imitation learning enables intelligent systems to acquire complex behaviors with minimal supervision. However, existing methods often focus on short-horizon skills, require large datasets, and struggle to solve long-horizon tasks or generalize across task variations and distribution shifts. We propose a novel neuro-symbolic framework that jointly learns continuous control policies and symbolic domain abstractions from a few skill demonstrations. Our method abstracts high-level task structures i…

Renormalization Group Running of CPT Operators In Lorentz-Violating Quantum Field Theory
Brett Altschul
https://arxiv.org/abs/2509.25488 https://arxiv.org/…

Renormalization Group Running of CPT Operators In Lorentz-Violating Quantum Field Theory
In conventional relativistic quantum field theory, the discrete operators C, P, and T are matrix operators with no renormalization scale dependence. However, in a Lorentz-violating theory with a fermion $f^μ$ term in the action, these operators may acquire nontrivial renormalization group behavior. Because the $f^μ$ term may actually be exchanged in the action for an equivalent $c^{νμ}$ term, the scale dependence depends explicitly on the renormalization scheme, even at one-loop order. The …

I just signed the campaign: Sauvons les trains de nuit Paris – Berlin et Paris – Vienne !
Save the night trains Berlin-Paris & Paris-Vienna !
For slow, lower carbon travel, it would be great if you could also add your name to this important issue. Every name that is added builds momentum around the campaign and makes it more likely for us to get the change we want to see.
Will you join me by taking action on this campaign?

🚂 Sauvons les trains de nuit Paris-Berlin et Paris-Vienne
Signez la pétition maintenant !

Americans Don't Agree with Democrats' Actions (President Donald J. Trump/The White House)
https://www.whitehouse.gov/government-shutdown-clock/
http://www.memeorandum.com/250930/p149#a250930p149

Kid2 had pizza day yesterday. The school decided to order from a new place that offers "healthy pizza." The kids apparently poured onto the playground after lunch, having barely touched the stuff, chanting "We want [corporate pizza chain]!" While I may not agree with their taste in pizza, I'll take collective action any day. And the kids won! The school sent around an email last night saying they were going back to the old place and that they'd pay for a pizza par…

Actions as Language: Fine-Tuning VLMs into VLAs Without Catastrophic Forgetting
Asher J. Hancock, Xindi Wu, Lihan Zha, Olga Russakovsky, Anirudha Majumdar
https://arxiv.org/abs/2509.22195

Actions as Language: Fine-Tuning VLMs into VLAs Without Catastrophic Forgetting
Fine-tuning vision-language models (VLMs) on robot teleoperation data to create vision-language-action (VLA) models is a promising paradigm for training generalist policies, but it suffers from a fundamental tradeoff: learning to produce actions often diminishes the VLM's foundational reasoning and multimodal understanding, hindering generalization to novel scenarios, instruction following, and semantic understanding. We argue that this catastrophic forgetting is due to a distribution mismatch …

How can Western countries back up Palestine recognition with action? Here are 4 ways to pressure Israel https://theconversation.com/how-can-western-countries-back-up-palestine-recognition-with-acti…

How can Western countries back up Palestine recognition with action? Here are 4 ways to pressure Israel
Western countries agree a two-state solution is the way to resolve the Israeli-Palestinian conflict. But are they willing to take action to make it happen?

Does Calibration Affect Human Actions?
Meir Nizri, Amos Azaria, Chirag Gupta, Noam Hazon
https://arxiv.org/abs/2508.18317 https://arxiv.org/pdf/2508.18317

Does Calibration Affect Human Actions?
Calibration has been proposed as a way to enhance the reliability and adoption of machine learning classifiers. We study a particular aspect of this proposal: how does calibrating a classification model affect the decisions made by non-expert humans consuming the model's predictions? We perform a Human-Computer-Interaction (HCI) experiment to ascertain the effect of calibration on (i) trust in the model, and (ii) the correlation between decisions and predictions. We also propose further correct…

GIT stability and biquotients of $SU(3)$
Yoshinori Hashimoto, Hiroaki Ishida, Hisashi Kasuya
https://arxiv.org/abs/2509.20254 https://arxiv.org/pdf/2509.20…

GIT stability and biquotients of $SU(3)$
We study double-sided actions of $(\mathbb{C}^*)^2$ on $SL(3,\mathbb{C})/U$ and the associated quotients, where $U$ is a maximal unipotent subgroup of $SL(3,\mathbb{C})$. The main results of this paper are a sufficient condition for the double-sided quotient to agree with the quotient in terms of the geometric invariant theory (GIT), and an explicit necessary and sufficient condition for $SL(3,\mathbb{C})/U$ to agree with the $χ$-stable locus in its affine closure. We apply this result to char…